The analyst firm cited the company’s first-quarter performance and wrote in the research note that eBay remains on course to see potential gross merchandise volume (GMV) growth during the second half of 2024. Recent product development and generative AI applications were highlighted as a way the company will improve productivity.

“The key to our AI strategy is adding GPU capacity that combines with our existing infrastructure to develop large language models in-house,” he said.

During the first three months of 2024, eBay doubled its graphics processing unit (GPU) capacity, compared with the end of 2023. GPU is a computer chip that can perform high-speed mathematical calculations required for generative AI (GAI) and large language models (LLMs).

Mekel-Bobrov said eBay continues to take a hybrid approach, leveraging both commercial and open-source LLMs, but expects more GAI services to be powered by the company’s internal LLM this year.

GAI powers consumer-facing features including Explore, Shop the Look, Magical Listing, and background enhancement, but increasingly the company leverages GAI to change how employees drive meaningful productivity and efficiency improvements across the organization.

It's interesting to think about the transformation -- how a company founded on September 3, 1995, in San Jose, California exploded in growth from a global online auction and trading company launched by tech entrepreneur Pierre Omidyar into a technology powerhouse.

Mekel-Bobrov pointed to its Global Customer Experience organization, GCX, as one of many transformative ventures the company is going through. He said eBay is in the midst of a multi-year journey to transform its GCX capabilities using technology.

“Our north star for customer service is not just to reduce costs, but meaningfully improve the quality of interactions,” he said.  

GAI tools also improve developer productivity and accelerate eBay’s overall tech velocity. Last year, eBay made GitHub Copilot available to all of its developers and saw “encouraging” results. “We observed measurable improvements in productivity along with code acceptance and accuracy,” he said. “Nearly three-quarters of our developers now use Copilot every day.”

It’s still early in eBay’s GAI productivity journey, but these few examples of proprietary tools  introduced to its GCX and technology teams are already yielding tangible benefits.